Unravelling programme success and complex causation in Agricultural Research for Development (AR4D): A systematic and comprehensive literature review

This article presents a systematic and comprehensive literature review of agricultural research for development (AR4D) programs. The review aims to distil commonly mentioned outcomes of AR4D programs and intervention, their causal conditions, and their causal relationships. The review also seeks to unpack what complex causation means in the context of AR4D. Following PRISMA guidelines, the review covers the period from 1980 to June 2023 and includes a meticulous selection of peer-reviewed journal articles, books, chapters, and grey literature (n = 57 from an initial sample of n = 427). The findings reveal a clear and limited set of outcomes and conditions, highlighting coherence and manageability within the field. However, concerns arise regarding the representation of real-world AR4D programs, with a bias towards reporting positive outcomes and successful initiatives while overlooking less successful or failing programs, particularly from Latin America and Central Asia. Complex causation emerges as a recurrent theme, emphasizing the need for innovative research methods to understand the intricate relationships between outcomes and multiple contributing factors. Furthermore, scaling successful programs is a pressing topic, challenging assumptions of replicability and calling for a comprehensive understanding of scaling processes. This study underscores calls in the field for adopting alternative research methods and critically evaluating the existing knowledge base to enhance the effectiveness and impact of AR4D programmes and interventions. [Display omitted] •AR4D review shows limited program outcomes and conditions, raising concerns about representation and regional bias.•AR4D programs involve complex causal relationships, necessitating innovative methods to understand outcome combinations.•Researchers advocate for process tracing, qualitative comparative analysis, and realist evaluations in AR4D studies.•The review highlights the need for a critical approach to scaling assumptions and a “science of scaling” in AR4D research.